Orderly and specific connections among neurons of multiple types underlie the function of neural circuits. Our work has used mouse retina as a model to identiy molecules and mechanisms that underlie this specificity. The retina is not only important and fascinating in its own right, but is also a particularly accessible portion of the central nervous system. Moreover, we and others have generated a set of transgenic lines that allow specific retinal cell types to be marked and manipulated. We focus on the inner plexiform layer (IPL) in which dozens of types of interneurons (amacrine and bipolar cells) form stereotyped patterns of connectivity with ~30 types of retinal ganglion cells (RGCs), endowing each of them with sensitivity to visual features such as motion in a particular direction or color contrast. Wiring up these largely hard-wired circuits seems likely to require many cell type-specific molecules. Over the past several years he have identified several candidates, and tested them in vivo using loss- and gain-of-function strategies. During the past project period, we found two roles for members of the cadherin superfamily of recognition molecules in assembly of the IPL. First, a pair of classical cadherins (Cdh8 and Cdh9) play instructive roles in directing axonal arbors of two bipolar cell types (BC2, BC5) to appropriate sublaminae within the IPL. Second, a group of clustered gamma protocadherins (Pcdhg's) are required for dendritic patterning of starburst amacrine cells (SACs) in the lateral plane. Fortuitously, BC2, BC5 and SACs are all components of a direction-selective circuit that also includes ON-OFF direction-selective RGCs, which send information about motion in four cardinal directions to the rest of the brain. These results provide us with the opportunity to address a long- standing issue: how do members of multigene families work together in combinations to pattern neural circuits. Using new genome editing methods, we have generated double and triple mutants that lack (a) both Cdh8 and its closest relative Cdh11, (b) Cdh9 and its closest relatives, Cdh6 and Cdh10, and (c) Pcdhgs and their closest relatives, the alpha-Pcdhs. Importantly all of these genes are expressed by cells of the direction-selective circuit. Our aims now are to analyze mutants in which genes of the cadherin superfamily have been mutated singly and in combination. We will use molecular, histological and electrophysiological methods to analyze the consequences of these perturbations on the structure and function of the direction-selective circuit. Together, these studies will allow us to take a first step toward confronting the disturbing reality that analysis of single genes is insufficient to understand the assembly of complex neural circuits.